Optimisation of planning and delivery methods for Gamma Knife Stereotactic Radiosurgery

The work presented in this dissertation is the response to a series of technical and clinical challenges encountered during the first four years of operations in the Gamma Knife Centre at Queen Square. These challenges, which are common to most contemporary Gamma Knife centres, were prospectively addressed and practical solutions were developed for the questions they posed. The dosimetric differences between the new convolution algorithm and the standard TMR 10 algorithm traditionally used for prediction of dose distribution in Gamma Knife Radiosurgery (GKR) were explored. It was demonstrated that inhomogeneity correction with the convolution algorithm results in a considerable but consistent dose shift compared to TMR 10. No significant difference in relative dose distribution was noted and it was concluded that a reduction of the prescription dose is necessary to obtain the same absolute dosimetric effect with the convolution algorithm. The stability of the stereotactic Leksell frame G in GKR was demonstrated using a comprehensive study design that involved repeated measurements of landmarks by two observers. The study provided reliable and realistic evidence of submillimetre stability of the stereotactic frame throughout the treatment procedure which is important for evaluation and development of new frameless radiosurgery systems. The technical feasibility of using a combination of three magnetic resonance angiography sequences (triple-MRA), instead of digital subtraction angiography (DSA), for visualisation and delineation of brain Arteriovenous Malformations (AVMs) for GKR targeting was demonstrated. Target volumes obtained using triple-MRA are on average 10% smaller than AVM targets obtained with the standard DSA planning method and this can potentially reduce the risk of adverse radiation effects (ARE). The treatment planning method described here has laid the way for a change in clinical practice that favours a less invasive treatment planning approach. The same principle of less invasive AVM imaging with triple-MRA was used at the post-GKR stage, when a DSA is performed to confirm AVM obliteration. Triple-MRA was found to consistently confirm or rule out residual AVMs in patients who had undergone GKR for brain AVMs, compared to DSA, and it can also be reliably used for characterisation of residual AVMs. The use of triple-MRA as a first line for assessment of obliteration after GKR is recommended